NO330262B1 - Procedure for in vitro molecules development of antibody effect, in vitro. - Google Patents

Abstract

The present invention relates to a method for in vivo molecular evolution of antibody function. According to the present invention, a nucleic acid encoding a CDR that is normally contained in a framework (the “original framework”), which differs from a selected master framework, is amplified from an immunoglobulin gene and is inserted into a nucleic acid encoding the selected master framework. The invention further provides an antibody library, such as a phage display library, and methods of making the same.

Description

The present invention relates to a method for molecular development of antibody action, in vitro.

Background

WO 98/32845, Soderlind et al (1999) and Jirholt et al (1998) describe in vitro molecular development of antibody-derived proteins, by implanting naturally occurring complementarity-determining regions (CDRs) into a defined and selected ("major ") frame, comprising the frame regions from the germline gene DP-47 in the VH3 family. Oligonucleotide primers, based on the sequences encoding portions of the framework regions of DP-47 immediately flanking the CDRs, are used to amplify nucleic acid sequences encoding CDRs from a DNA library prepared by peripheral blood B cells. Single-stranded DNA from the amplification reaction is then combined with overlapping oligonucleotides encoding the remainder of the main frame in an overlap-expansion PCR reaction to produce full-length sequences encoding VH antibody regions, each of which contains in-frame regions of the DP-47 germline gene, and CDRs from the germline genes in the VH3 family.

This technique provides a valuable tool for increasing the diversity of antibody libraries (eg, phage display libraries), particularly as it allows recombination of CDR1 and CDR2, which are usually joined in vivo and do not undergo recombination during germline gene rearrangement .

Furthermore, the CDRs have been proofread in vivo and it is unlikely that these are immunogenic, which gives an advantage compared to artificially mutated CDR sequences. The main frames used (from germ cell line DP-47 and DPL3) have also been chosen to be highly compatible with the bacterial expression and phage system used, and a high degree of functional protein display is therefore ensured.

Another important aspect of building up the library was a requirement for the frame to be able to maintain specificities with high affinity, against a wide range of different types of antigens. As the system allows variability to be introduced in any number of the CDR positions, the variability achieved is great, far beyond what can be achieved with previously established combination technologies. Finally, the modular design, i.e. the fact that variability is introduced into a common frame structure, makes subsequent modifications and studies of selected clones easy and efficient.

Using this technology, tentatively called CDR implantation, a large phage-display antibody library based on the single-stranded Fv (scFv) format has been generated. The library is used to select a panel of high-affinity antibodies against a variety of ligand types, including proteins (of human and non-human origin), peptides, carbohydrates and low molecular weight haptens. From a functional point of view, it is therefore clear that a simple, selected backbone can hold antibody specificities with high affinity against quite different types of antigens, which suggests that the topology on the surfaces can vary widely between the antibodies.

However, there remains a general need within the field to increase the diversity of the antibody libraries.

The invention

In accordance with the present invention, nucleic acids encoding a CDR usually found in a frame (the "original frame"), which is different from a selected main frame, are amplified from an immunoglobulin gene and inserted into a nucleic acid encoding the selected main frame.

Amplification can be carried out as for normal CDR implantations as described above, by PCR using primers based on frame regions flanking the CDRs. In the present invention, however, there is a difference between the original frame and the main frame. In contrast to previously described CDR implantation methods, nucleic acids encoding CDRs are therefore not amplified using primers based exclusively on the main frame. Rather, primers are used that differ from the sequence that codes for the main frame. The primers can be based on a specifically given other frame, e.g. from another germline gene, or a consensus sequence of a germline gene family, or they may be degenerate, e.g. to amplify CDRs from a variety of germline families.

In a first aspect, the present invention therefore provides a method for the production of a polynucleotide sequence that codes for an antibody variable region, the variable region comprising complementary determining regions (CDRs) located within a selected frame ("master frame"), the method comprises the steps: a) providing at least one nucleic acid molecule encoding one or more CDRs and associated framework regions (the original framework); b) amplification of at least one CDR-encoding part of the nucleic acid molecule(s) in step a) using one or more pairs of oligonucleotides as amplification primers, and c) assembly of a polynucleotide sequence encoding an antibody -variable region by combining the amplified CDR coding nucleotide sequences produced in step b) with

nucleic acid sequences that encode the main frame,

in that the oligonucleotide primers from step b) comprise nucleotide sequences that are different from the corresponding nucleotide sequences that code for the main frame.

Another aspect of the invention provides a method for the production of a library of polynucleotide sequences each encoding an antibody variable region comprising complementary determining regions (CDRs) located within a common selected frame (master frame), the method comprising the steps: a) providing a population of nucleic acid molecule encoding one or more complementary determining regions (CDRs) and associated framework regions (the original framework); b) amplification of at least one CDR-encoding part of the nucleic acid molecule(s) in step a) using one or more pairs of oligonucleotides as amplification primers, and c) assembly of a polynucleotide sequence encoding an antibody -variable region by combining the amplified CDR-encoding nucleotide sequences produced in step b) with

nucleic acid sequences that encode the main frame,

in that the oligonucleotide primers from step b) comprise nucleotide sequences that are different from the corresponding nucleotide sequences that code for the main frame.

By "belonging framework regions" as used in relation to the nucleic acid molecules obtained in step a), amino acid residues of the framework region which immediately flank the CDR are meant. For example, the nucleic acid molecule(s) in step a) may encode a CDR together with up to 5, 10, 15 or more in-frame amino acid residues flanking one of the sides of the CDR. In this way, the nucleic acid molecule(s) may encode an antibody variable region or even an entire antibody.

By "different from" in connection with nucleotide sequences of the oligonucleotide primers in step b) it is meant that the regions of the oligonucleotide primers in step b) which code for frame residues (i.e. those regions of the oligonucleotide primers which are complementary to regions on the nucleic acid molecules obtained in step a) which encode frame residues), do not have 100% sequence identity with the corresponding regions on the nucleic acid molecules which encode the main frame.

In a preferred embodiment of the first and second aspects of the invention, the method comprises the steps: i) providing at least one pair of oligonucleotides;

ii) using each said pair of oligonucleotides as amplification primers to amplify nucleotide sequences encoding different CDRs, and

iii) assembly of polynucleotide sequences encoding antibody variable regions by incorporating nucleotide sequences derived from step ii) above with nucleotide sequences encoding framework sequences (FRs) of a selected type;

wherein the oligonucleotides in step (i) have sequences which are different from the corresponding sequences which code for the said main frame.

For the avoidance of doubt, the "frame" of a variable region, as used herein, is typically formed by up to four individual framework regions, flanking the three CDRs of the variable region:

The "frame regions (FRs) of a selected type" together provide a "main frame".

It will be understood by those skilled in the art that the methods in accordance with the present invention can be used to produce a polynucleotide sequence which codes for a variable region of different types of antibody. For example, the variable region can be an IgG, IgM, IgA, IgD or IgE variable region.

The polynucleotide sequence assembled in step c) preferably codes for an IgG variable region. Preferably, the polynucleotide sequence(s) assembled in step c) encodes an IgG heavy chain or light chain. It is appropriate if the polynucleotide sequence(s) assembled in step c) encodes an unnaturally occurring antibody variable region.

The polynucleotide sequence(s) assembled in step c) preferably encodes an antibody variable region comprising at least one CDR having a canonical structure atypical of antibody variable regions comprising the main frame. By "atypical" is meant that the CDR has a canonical structure found in less than 10% of the naturally occurring antibody variable regions comprising the selected main frame. The CDR preferably has a canonical structure found in less than 5%, 2% or 1% of the naturally occurring antibody variable regions comprising the selected main frame. More preferably, the CDR has a canonical structure that is not found in any naturally occurring antibody variable regions comprising the selected main frame.

In a preferred embodiment of the methods according to the invention, at least one of the polynucleotide sequence(s) assembled in step c) codes for an antibody variable region

comprising at least one CDR derived in a different germline gene family from that of the main frame. For example, the polynucleotide sequence(s) assembled in step c) may encode an antibody variable region comprising one or more CDRs derived from a light chain in a frame of a heavy chain, or vice versa.

Step a) preferably comprises obtaining a population of nucleic acid molecules each encoding an antibody variable region. The nucleic acid molecules can advantageously each code for an antibody variable region from the same germline gene family. Selectivity for the CDRs to be incorporated into the main frame can thus be achieved, and at least in part, by providing a selected population of nucleic acid molecules in step a).

Alternatively, or in addition, selectivity for the CDRs to be incorporated into the main frame can be achieved by using oligonucleotide primer pairs in step b) that selectively hybridize with a targeted sub-population of nucleic acid molecules obtained in step a). By "selective hybridization" it is meant that the oligonucleotide primer pairs hybridize selectively with a targeted sub-population of the nucleic acid molecules under stringent conditions. Oligonucleotide hybridization conditions are described in Molecular Cloning: A Laboratory Manual (Third Edition), Sambrook and Russell (eds.), Cold Spring Harbor, Laboratory Press.

It will be understood by those skilled in the art that the ability of the primers to selectively hybridize with target nucleic acid molecules largely depends on the degree of sequence complementarity between the primers and the target sequences.

It is advantageous that the oligonucleotide primer pairs in step b) hybridize selectively with a targeted sub-population of nucleic acid molecules obtained in step a), each of which encodes an antibody variable region from the same germline gene family.

By obtaining a selected population of nucleic acid molecules in step a) and/or by using oligonucleotide primer pairs in step b) which selectively hybridize with a targeted sub-population of the nucleic acid molecules obtained in step a), it is on in this way possible to select the CDRs to be incorporated into the main frame.

In a preferred embodiment, the CDRs to be incorporated into the main frame are derived from the nucleic acid molecules encoding an antibody variable region from the same germline gene family, such as the VH3 family. The CDRs to be incorporated into the main frame are advantageously derived from the nucleic acid molecules encoding an antibody variable region from the same germline gene, such as DP-29 and DP-73.

The main frame is advantageously derived from a germline gene selected from the group comprising DP-47 and DPL-3.

In a preferred embodiment of the methods according to the invention, step c) comprises the use of overlap expansion PCR (see Sambrook and Russell, supra). In this case, it is necessary to isolate the single-stranded nucleic acid molecule from the amplified CDR-encoding nucleic acid molecules produced in step b). This can be achieved by using oligonucleotide primer pairs where one of the primers is biotinylated, and thus the nucleic acid strand produced by extension of the biotinylated primer can be isolated on the basis of its affinity for streptavidin. The use of biotinylated primers and overlap expansion PCR is described in Jirholt et al., 1998, supra.

As a consequence of the differences in the nucleic acid sequence between the regions of the amplification primers encoding frame residues and nucleic acid sequences encoding the corresponding regions of the main frame, the amplified CDRs cannot always be incorporated directly into the main frame, by overlap expansion PCR , as the nucleotides that code for the amplified CDRs do not anneal sufficiently with the nucleic acid that codes for the main frame. This is particularly the case when (as in example IB) there are significant misfits

(eng. mismatches) at the end of the amplification primer (especially at the end closest to the CDR). As a result, it may be necessary to change the regions of the amplified nucleotide sequences that encode the framework regions flanking the CDRs, to make them more similar in sequence to the regions of nucleic acid molecules that encode the corresponding regions of the main framework.

In a preferred embodiment of the methods in accordance with the invention, the methods therefore comprise a further step, carried out after step b) and before step c), for modifying the nucleotide sequence of the amplified CDR-encoding molecules in step b) so that the parts of the amplified molecules encoding in-frame regions have greater sequence identity with the corresponding parts of the main frame. The nucleotide sequences are preferably modified so that they have 100% sequence identity with the corresponding parts of the main frame.

One way of doing this is, for example, to initially amplify the CDR and adjacent parts of the flanking frame regions using PCR primers that are identical or very similar to the original frame of the CDR. Then, in subsequent rounds of PCR amplification, the regions of the amplified nucleic acid sequences that code for frame regions can be modified, using primers that contain mismatches in relation to the original frame, the mismatches producing the residues that are present on the corresponding positions in the main frame. Alternatively, the in-frame regions can be sufficiently similar, or identical, to the main frame to be incorporated using overlap expansion PCR.

Alternatively, a single round of PCR amplification using primers representing a chimera of the original frame and main frame may be sufficient to amplify CDRs that can be used in the overlap expansion PCR process, in which case the additional step will not be necessary.

In the first round of PCR (step b), it may therefore be advantageous, regardless of whether said additional step is to be performed, to include mismatches in the primers relative to the original frame, in order to include as many bases as possible that are in common with it main frame selected. The number of mismatches that can be included depends on a number of factors including the number of bases that differ between the frames, the length of the primers, and the risk of pairing with sequences other than those intended.

It may be possible for CDRs amplified by primers identical to the original frame to be incorporated directly into the main frame (ie the product of step b) by overlap expansion PCR. Alternatively, additional PCR steps (as described above) may be required so that the frame sequences associated with the amplified CDR encode more closely to the corresponding sequences in the main frame.

In a preferred embodiment of the first aspect of the invention, the method comprises a further step of inserting the polynucleotide sequence(s) assembled in step c) into an expression vector. The expression vector is preferably a secretion vector.

The polynucleotide sequences produced by the methods of the invention can thus be used in accordance with known techniques, appropriately modified with respect to the teachings provided herein, to construct an expression vector, which is then used to transform a suitable host cell for expression and production of antibody variable regions. Such techniques include those described in US Patent 4,440,859, issued April 3, 1984 to Rutter et al, 4,530,901 issued July 23, 1985 to Weissmann, 4,582,800 issued April 15, 1986 to Crowl, 4,677,063 issued June 30, 1987 to Mark et al, 4,678,751 issued 7 July 1987 to Goeddel, 4,704,362 issued November 3, 1987 to Itakura et al, 4,710,463 issued December 1, 1987 to Murray, 4,757,006 issued July 12, 1988 to Toole, Jr. et al, 4,766,075 issued August 23, 1988 to Goeddel et al and 4,810,648 issued March 7, 1989 to Stalker.

The polynucleotide sequences produced by methods in accordance with the invention can be linked to a wide range of other DNA sequences for introduction into a suitable host. Companion DNA will depend on the nature of the host, how the DNA is introduced into the host, and whether episomal maintenance or integration is desired.

Typically, the polynucleotide sequence is inserted into an expression vector, such as a plasmid, in the correct orientation and with the correct reading frame for expression. If necessary, the polynucleotide sequence can be linked to suitable transcriptional and translational regulatory nucleotide sequences recognized by the desired host, although such controls are usually available in the expression vector. The polynucleotide sequence insert can in this way be operatively linked to a suitable promoter. Bacterial promoters include the E. coli lad and lacZ promoters, T3 and T7 promoters, the gpt promoter, phage A, the PR and PL promoters, the phoA promoter and the trp promoter. Eukaryotic promoters include the immediate early CMV promoter, the HSV thymidine kinase promoter, the early and late SV40 promoter, and the promoters of retroviral LTRs. Other suitable promoters will be known to those skilled in the art. It is desirable that the expression constructs also contain sites for transcription initiation and termination, and in the transcription region, a ribosome binding site for translation (Hastings et al., International patent no. WO 98/16643, published 23 April 1998 ).

The vector is then introduced into the host using standard techniques. Typically, not all hosts will be transformed by the vector, and it will therefore be necessary to select the transformed host cells. A selection technique involves the incorporation of a DNA sequence marker with any necessary control element, encoding a selectable trait in the transformed cell, into the expression vector. Such markers include dihydrofolate reductase, G418 or neomycin resistance for eukaryotic cell cultures, and tetracycline, kanamycin or ampicillin resistance genes for cultivation in E. coli and other bacteria. Alternatively, the gene for such a selectable trait can be another vector, which is used to co-transform the desired cells.

Host cells transformed by the expression vector are then cultured for a sufficient time and under suitable conditions known to those skilled in the art in light of the teachings provided herein to allow expression of the encoded antibody variable region, which can then be recovered.

The antibody variable region can be recovered and purified from recombinant cell cultures by well-known methods, including ammonium sulfate or ethanol precipitation, acid extraction, anion or cation exchange chromatography, phosphocellulose chromatography, hydrophobic interaction chromatography, affinity chromatography, hydroxylapatite chromatography and lectin chromatography. High performance liquid chromatography ("HPLC") is preferably used for purification.

Many expression systems are known, including systems that use: bacteria (eg, E. coli and Bacillus subtilis) transformed with, for example, recombinant bacteriophage, plasmid or cosmid DNA expression vectors; yeast (eg, Saccaromyces cerevisiae) transformed with, for example, yeast expression vectors; insect cell systems transformed with, for example, viral expression vectors (eg, baculovirus), plant cell systems transfected with, for example, viral or bacterial expression systems, animal cell systems transfected with, for example, adenovirus expression vectors.

The vectors may include a prokaryotic replicon, such as Col El ori for propagation in a prokaryote, although the vector is to be used for expression in other non-prokaryotic cell types. The vectors can also include a suitable promoter such as a prokaryotic promoter which can direct the expression (transcription and translation) of genes in a bacterial host cell, such as E. coli, transformed with it.

A promoter is an expression control element formed by a DNA sequence that allows binding of RNA polymerase and transcription. Promoter sequences compatible with exemplary bacterial hosts are usually provided in plasmid vectors containing appropriate restriction sites for insertion of a DNA segment in accordance with the present invention.

Common prokaryotic vector plasmids are: pUC18, pUC19, pBR322 and pBR329 available from Biorad Laboratories (Richmond, CA, USA), p7rc99A, pKK223-3, pKK233-3, pDR540 and pRIT5 available from Pharmacia (Piscataway, NJ; USA), pBS vectors , Phagescript vectors, Bluescript vectors, pNH8A, pNH16A, pNH18A, pNH46A available from Stratagene Cloning Systems (La Jolla, CA 92037, USA).

A common mammalian cell vector plasmid is pSVL from Pharmacia (Piscataway, NJ, USA). This vector uses the SV40 late promoter to drive expression of cloned genes, the highest level of expression being found in T antigen-producing cells such as COS-1 cells. An example of an inducible mammalian expression vector is pMSG, also available from Pharmacia (Piscataway, NJ, USA). This vector utilizes the mouse mammary tumor virus long terminal repeat glucocorticoid-inducible promoter to drive expression of the cloned gene.

Useful yeast plasmid vectors are pRS403-406 and pRS413-416 and are usually available from Stratagene Cloning Systems (La Jolla, CA 92037, USA). Plasmids pRS403, pRS404, pRS405 and pRS406 are Yeast Integrating plasmids (Ylps) and incorporate the yeast selectable markers HIS3, TRP1, LEU2 and URA3. Plasmids pRS413-416 are Yeast Centromere plasmids (YCp's).

Methods well known to those skilled in the art can be used to construct expression vectors containing the coding sequence and, for example, appropriate transcriptional and translational controls. Such a procedure involves ligation via homopolymeric tails. Homopolymer polydA (or polydC) tails are added to exposed 3'OH groups on the DNA fragment to be cloned, by terminal deoxynucleotide transferases. The fragment is then able to ligate with polydT (or polydG) tails attached to the ends of a linearized plasmid vector. The gaps after the following joining can be filled with DNA polymerase and the free ends joined with DNA ligase.

Another method involves ligation via cohesive ends. Compatible cohesive ends can be generated on the DNA fragment and the vector by suitable restriction enzymes. These ends will quickly join together through complementary base-pairing and the remaining incision can be closed by DNA ligase.

A further method uses synthetic molecules called linkers and adapters. DNA fragments with blunt ends are generated by the bacteriophage T4 DNA polymerase, or E. coli DNA polymerase I, which removes protruding 3' ends and fills in retracted 3' ends. Synthetic linkers, parts of blunt-ended double-stranded DNA containing recognition sequences for defined restriction enzymes, can be ligated with blunt-ended DNA fragments with T4 DNA ligase. They are then degraded with suitable restriction enzymes to form cohesive ends, and ligated into an expression vector with compatible ends. Adapters are also chemically synthesized DNA fragments that contain a blunt end used for ligation, but also contain a pre-formed cohesive end.

Synthetic linkers containing a variety of restriction endonuclease sites are commercially available from a number of sources, including International Biotechnologies Inc, New Haven, CN, USA.

A desired way to modify the DNA encoding the polypeptide in accordance with the invention is to use the polymerase chain reaction as discussed by Saiki et al (1988) Science 239, 487-491.1 this method the DNA to be amplified enzymatically flanked by two specific oligonucleotide primers which are themselves incorporated into the amplified DNA. The specific primers mentioned may contain restriction endonuclease recognition sites which can be used for cloning into expression vectors, using methods known in the art.

A third aspect of the present invention therefore provides a polynucleotide sequence which can be produced and/or produced by a method in accordance with the first or second aspect of the invention.

A fourth aspect of the invention provides a polypeptide encoded in a polynucleotide sequence in accordance with the third aspect of the invention, for example an antibody or a fragment thereof (eg single-stranded ScFv antibody).

The invention further provides a vector incorporating a polynucleotide sequence in accordance with the second aspect of the present invention, and host cells transformed with such vectors. Exemplary host cells include mammalian cells such as Chinese hamster ovary cells.

In a preferred embodiment, the expression vector is a phage display vector.

The display of proteins and polypeptides on the surface of bacteriophage (phage), linked together with one of the phage coat proteins, provides a powerful tool for the selection of specific ligands. This "phage display" technique was originally used by Smith in 1985 (Science 228, 1315-7), to generate large antibody libraries for the purpose of selecting those with high affinity for a given antigen. Recently, the method has been used to present peptides, regions of proteins and intact proteins on the surface of phage to identify ligands that have the desired properties.

The principles behind subject display technology are as follows:

(i) A nucleic acid encoding the protein or polypeptide for display is cloned into a phage; (ii) The cloned nucleic acid is expressed and fused with the coat-anchoring part of one of the phage coat proteins (usually the p3 or p8 coat protein in the case of a filamentous phage), so that the foreign the protein or polypeptide is displayed on the surface of the subject; (iii) The subject that displays the protein or polypeptide with the desired properties is then selected (e.g. by affinity chromatography), thus providing a genotype (linked to a phenotype) that can be sequenced, multiplied and transferred to other expression systems .

Alternatively, the foreign protein or polypeptide can be expressed using a phagemid vector (ie a vector comprising the origin of replication derived from a phage and a plasmid) which can be packaged as a single-stranded nucleic acid in a bacteriophage coat. When the plasmid vectors are used, a "helper phage" is used to provide the functions of replication and packaging of the phagemid nucleic acid. The resulting phage will express both the wild-type coat protein (encoded for by the helper phage) and the modified coat protein (encoded for by the phage mite), while only the modified coat protein is expressed when a phage vector is used.

Methods for selecting subjects that express a protein or peptide with a desired specificity are known in the art. For example, a widely used procedure is "panning", whereby phage stocks displaying ligands are exposed to solid-phase linked target molecules, e.g. using affinity chromatography.

Alternative procedures for the selection of relevant phages include SAP (Selection and Amplification of Phages, as described in WO 95/16027) and SIP (Selectively-lnfective Phages, EP 614989 A, WO 99/07842) which use selection based on amplification of phages, whereby the presented ligand is bound specifically to a ligand binder. In one embodiment of the SAP method, this is achieved by using non-infectious phages and by connecting the appropriate ligand linker to the N-terminal part of p3. If the ligand binder is specifically bound to the displayed ligand, in this way the otherwise non-infectious ligand-expressing phage is supplied with the parts of p3 that are necessary for infection. As this interaction is reversible, selection can thus be based on kinetic parameters (see Duenas et al., 1996, Mol. Immunol. 33, 279-285).

The use of phage display to isolate ligands that bind biologically relevant molecules is reviewed in Felici et al., (1995) Biotechnol. Annual. Fox. 1,149-183, Katz (1997) Annual Rev. Biophys. Biomol. Struct. 26, 27-45 and Hoogenboom et al. (1998) Immunotechnology 4(1), 1-20. Several randomized combinatorial peptide libraries have been designed to select polypeptides that bind different targets, e.g. cell surface receptors or DNA (reviewed by Kay, 1995, Perspect. Drug Discovery Des. 2, 251-268; Kay and Paul, 1996, Mol. Divers. 1139-140). Proteins and multimeric proteins have been successfully phage-displayed as functional molecules (see EP 0349578A, EP 0527839 A, EP 0589877 A; Chiswell and McCafferty, 1992, Trends in Biotechnol. 10, 80-84). In addition, functional antibody fragments (eg, Fab, single-stranded Fv [scFv]) have been expressed (McCafferty et al., 1990, Nature 348, 552-554; Barbas et al., 1991, Proe. Nat. Acad. Sci. USA 88, 7978-7982; Clackson et al., 1991, Nature 352, 624-628) and some of the weaknesses of the human monoclonal antibody technology have been superseded as human high-affinity antibody fragments have been isolated (Marks et al., 1991, J. Mol. Biol, 222, 581-597; Hoogenboom and Winter, 1992, J. Mol. Biol. 227, 381-388). Further information on the principles and procedure of phage display is discussed in Phage display of peptides and proteins: a laboratory manual Ed Kay, Winter and McCafferty (1996) Academic Press, Inc ISBN 0-12-402380-0.

In a preferred embodiment of the first and second aspects of the invention, the method therefore further comprises the step of expressing the polynucleotide sequence(s) assembled in step c) and screening the resulting polypeptide(s) comprising an antibody- variable range, desired characteristics. The desired properties are preferably easily selectable with known techniques. For example, antibody can be screened for the desired affinity using affinity chromatography methods.

A further aspect of the invention provides a polynucleotide library which can be produced by a method in accordance with a second aspect of the invention, i.e. comprising polynucleotides which can be produced by a method in accordance with a first aspect of the invention. Such a library would comprise polynucleotides encoding a population of antibody variable regions, each of which shares a common framework ("master frame").

Preferably, the polynucleotide library is an expression vector library. It is appropriate if the polynucleotide library is a phage display library.

The invention

The present invention represents a development of the technology presented in WO 98/32845, Soderlind et al., (1999) Immunotechnology, 4, 279-285, and Jirholt et al. (1998) Gene 215, 471-476. Reference is made to these for a general description of the procedures and conditions used to amplify CDRs from a cDNA library containing antibody-encoding sequences, and procedures, materials and conditions for reassembling the CDRs thus amplified into the main frame by overlap-expansion PCR.

The method may further comprise the step of expressing the resulting antibody that the assembled nucleotide sequence encodes, and screening for desired properties. Again, this is described in detail in the above mentioned references.

The resulting expressed antibody can be screened for desired properties. For example, it may be desirable to change its ability to specifically bind to an antigen or to improve its binding properties relative to the parent antibody. Once again, this is described in detail in the above-mentioned references.

The oligonucleotides used as the amplification primers preferably have at least two nucleic acid residues which are different from a corresponding part of the nucleic acid sequence which codes for the main frame. More advantageously there are at least 3, 4, 5, 6, 7, 8, 10 or 12 different nucleic acid residues. In an alternative definition, the amplification primers preferably have no more than about 95% sequence identity with a corresponding portion of the nucleic acid sequence encoding the main frame, more advantageously no more than about 90%, 85%, 80%, 70% or 60% sequence identity.

In traditional CDR implantation, the amplification primers may include a small number of nucleotides encoding one or more amino acid residues of the adjacent end of the CDR (eg, three nucleotides, encoding one CDR residue). This also applies to the present invention, and in such cases the nucleotides of the CDR can be disregarded when the number of nucleotide differences between the primer and main frame is determined.

In light of the teachings provided herein and in the cited references, of basic CDR implantation techniques, one skilled in the art will be able to design primers to amplify the CDRs and, if necessary or desired, modify the amplification products to make frame regions more similar to the selected main frame.

When a given germline gene is to be targeted, highly specific primers may be desirable, for example, based precisely on the sequence encoding the portions of the framework regions of that gene flanking the CDR or CDRs to be amplified. The sequences of various germline genes are available from the VBASE sequence catalog (URL: http://www.mrc-cpe.cam.ac.uk/imt-doc/public/lNTRO.html) or from the DNAplot catalog (URL: http ://www.genetik.uni-koeln.de:80/dnaplot/vsearch_human.html).

Similarly, primers can be designed to amplify CDRs from a given germline gene family, by designing primers based on the consensus sequence of the genes in that family. For example, a consensus sequence may be defined as the base sequence found at >90% of the locus of a given germline family. Such sequences may include degeneracy sites, which suggest that different individual sequences have different nucleotides at that site. There may nevertheless be some common features of nucleotide residues that appear at such degeneracy sites, such sites are designated R (purine; bases G and A), Y (pyrimidine; C and T), M (amino; A, C), K (keto, T, G), S (strong; C, G), W (weak, A, T), B (not A), D (not C), H (not G) or V ( not T). A seat where no common feature is evident is marked N (if any).

Primers based on consensus sequences that include such designations may be degenerate, i.e. a population of primers is formed to include all possible combinations that are consistent with the consensus sequence, or where suitable artificial bases mimicking given sets of bases may be included in a homogeneous population of primers.

Information assigning germline genes to germline gene families (such as the variable heavy germline gene families VH1, VH2, VH3, VH4, VH5, VH6, and VH7) is available from the VBASE catalog referenced above.

Likewise, it is possible to design the primers to amplify CDRs from a variety of germline gene families, using a consensus sequence of germline genes from a variety of families. However, it will generally be preferred to target a particular germline gene or family.

With this in mind, a person skilled in the art will be able to design suitable primers depending on the specificity required. Preferably, at least one primer of the or each pair(s) used to initially amplify the CDRs is at least 15 nucleotides long, more preferably at least 18, even more preferably at least 21 or 24, optionally at least 30, 36 or 42. However, the primer is no more than 42 nucleotides long, more preferably no more than

36 or 30, more advantageously no more than 27.

The method will preferably be used to implant CDRs in all three positions in the variable region, as this leads to maximum variability, and ultimately more useful libraries. The method can, if desired (eg to optimize a previously obtained antibody), be used to implant only one or two CDRs. In such cases, the nucleic acid encoding the invariance CDR will be included in the overlap expansion PCR step, in addition to the newly amplified CDRs and the nucleic acid encoding the selected main frame.

A preferred embodiment of the invention is to implant CDRs from immunoglobulin genes of the same general type as the main frame (e.g. implanting VHCDRs in the VH main frame). However, the invention also encompasses the implantation of a CDR encoding nucleic acid from any type of immunoglobulin gene having a variable region as defined above, into a main frame which is independent of any such type of immunoglobulin superfamily gene. For example, VA CDRs can be inserted into a VH mainframe, and vice versa. Moreover, any member of the immunoglobulin superfamily having analogous structures with CDRs and FRs can provide the CDRs and/or main FRs in accordance with the invention, the above description being applicable mutatis mutandis.

The term "antibody" is used herein in its broadest sense, also to include variable region antibody fragments that include CDRs flanked by framework regions. Examples of antibody fragments which have such variable regions are the Fab fragment which comprises Vl, Vh, Clog CH1 regions; Fd fragment comprising VH and CH1 regions; Fv fragment comprising VLog VH regions of a single arm of an antibody; the dAb fragment comprising a VH region; and the F(ab')2 fragment, a bivalent fragment comprising two Fab fragments joined by a disulfide bridge at the hinge region. Single-stranded Fv fragments are also included.

Any desired main frame regions (or "frame regions (FRs) of a selected type") may be used in the present invention. In particular, they can be chosen to be highly compatible with the bacterial expression system and the phage system to be used, thus ensuring a high degree of functional protein display. Preferred examples are framework regions from the DP-47 and DPL-3 germline genes (of the VH3 and VÅ. germline gene families, respectively).

It is now generally agreed that CDR loops, which build up the surface of the antibody combination sites, can be grouped into a limited number of so-called canonical structures, depending on their conformation after folding. Pioneering work in this area was carried out by Cothia and Lesk (1987) who classified CDR1 and 2 of the heavy chain, and CDRs 1-3 of the light chain, into few basic structures.

The concept of canonical structures is the result of extensive analyzes of empirically determined and analyzed antibody structures. The determinants of canonical conformations are the length of the loops, key residues in the loops and key residues in the adjacent framework sequences (Chothia et al. 1992; Tomlinson et al. 1995; Al-Lazikani et al. 1997). For example, the human V* sequences can be grouped into 6 canonical structures for the CDRL1 loop, 1 canonical structure for the CDRL2 loop, and 5 canonical structures for the CDRL3 loop. In the same way, the human VH sequences can be grouped into 3 canonical structures for the CDRH1 loop and 4 canonical structures for the CDRH2 loop.

The CDRH3 loop has not yet been classified into distinct canonical classes, most likely due to its natural length variation leading to unique properties with respect to flexibility. Recently, however, it has been demonstrated that this CDR is also made up of structural elements that form a base torso, closely and to some extent inclusive of the frame region, and a top-head region that occasionally includes a further shoulder ( Morea et al. 1998).

It is conceivable that nature has developed different types of canonical structures to handle the diverse antigenic structures that the immune systems may encounter. Nature also presents these structures in conjunction with different framework structures. A given CDR loop is thus found in combination with a given frame (VBASE). There also appears to be a tendency for canonical structures to be used to form suitable surfaces, complementary to different types of antigens. In particular, it appears that loops with a canonical structure that build up a flat surface bind well to large protein antigens. These loops tend to be rather short, while longer loops are preferentially found in antibodies specific for smaller molecules, e.g. haptens (Lara-Ochoa et al. 1996). Not all the loops appear to be equally important in forming variability in the surfaces. H3 is of course particularly important in this respect, but also H2 and LI largely determine the surface area (Vargas-Madrazo et al. 1995).

Using CDR implantation technology, it has been unexpectedly discovered that some of the selected antibodies included CDRs with canonical structures that are not usually found in the frame used. These antibodies are functional as they bind the antigen with high affinity (Example 1). By using a single framework, it is thus possible to form functional variability in antibody combination sites that are based on canonical loops that are unusual in connection with a given framework. A library based on such a concept would have advantages over more conventional libraries as it can accommodate antibodies with a large variation in topologies and at the same time be highly efficient in the chosen host system (e.g. E. coli).

The binding properties of antibodies could also be improved by shuffling selected CDRs in order to be able to recombine the most optimal CDRs into a single antibody molecule. As will be appreciated, CDR implantation technology allows the generation of 1 to 6 CDRs simultaneously and has been used on the basis of the library presented herein in the examples to improve affinities of selected antibodies more than 30-fold in a single step.

The present invention can therefore lead to new combinations of classes of canonical structure, for example by combining canonical structures of classes that are not usually found in genes of the combined germline family. For example, by incorporating CDRH2 CDRs into the CDRL2 position in a VK chain, variability from 4 classes of canonical structure is available at this position, whereas in the natural VK antibody only one class of canonical structure is used in the CDRL2 position.

The amplification primers are preferably designed to amplify CDRs of a greater number of classes of canonical structure than the number of classes of canonical structure found in the germline gene family to which the main frame belongs, or CDRs of other classes of canonical structure than those found in the germline gene family to which the main frame belongs. Predictions of the canonical structure that a given CDR occupies can be determined using an online tool at URL: http://www.biochem.ucl.ac.uk/~martin/abs/chothia.html.

The master frame need not be one that occurs naturally, but may, for example, have been optimized, for example, for the expression or phage system to be used, or to reduce antigenicity in vivo.

The CDRs which have been amplified may be subjected to mutagenesis, for example using error-prone PCR, before being incorporated into the main frame (eg as described in WO 98/32845), although this is usually not preferred as naturally occurring CDRs are less likely to be antigenic than artificial ones.

"Percent (%) nucleic acid sequence identity" is defined as a percentage of the nucleic acid residues in a candidate sequence that are identical to the nucleic acid residues in the sequence to which it is compared, after aligning the sequences and gaps introduced, if necessary, to achieve the maximum percent sequence identity, and without considering any conservative substitutions as part of the sequence identity. The percent identity values given here were generated by the BLASTN module of WU-BLAST-2 (which was obtained from Altschul et al (1996);

URL:http://blast.wustl/edu/blast/README.html).WU-BLAST-2 uses several search parameters, most of which are set to default values. The adjustable parameters are set with the following values: overlap span = 1, overlap fraction = 0.125. A percent nucleic acid sequence identity value is determined by the number of matching identical residues divided by the total number of residues of the "longest" sequence in the aligned region, multiplied by 100. The "longest" sequence is the one with the most actual residues in the aligned region (gaps introduced by WU-BLAST-2 to maximize alignment judgment are ignored).

The following examples have been provided for a better understanding of the invention, and refer to the attached figures, where

figure 1 (parts A to D) shows the incorporation of a CDRH2 loop from the germ cell line gene DP29 into a frame of DP-47, and

figure 2 (parts A to E) shows the incorporation of a CDRH2 loop from the germ cell line gene DP-73 into

in a frame of DP-47.

Examples Designing primers and assembly of antibody genes

Primers different from the corresponding sequences in the DP-47 frame are used to amplify CDRs from different germline genes. In example IA, the main frame and the frame of the gene from which the CDR is amplified (DP-29) are sufficiently similar that the sequence amplified in this way can be incorporated into a DP-47 frame without further modification. In example IB, the frames are more different, and the sequence amplified in this way is further modified to be more similar to the DP-47 frame before being incorporated therein. This is accomplished by the use of primers that successively bring the in-frame regions flanking the CDRs into conformation with the selected frame in a designed and planned iterative process. In this way, it becomes possible to pick up CDR loops that have canonical structures that are atypical of the selected DP-47 framework.

When, as here, it is desirable to incorporate a specific CDR into the main frame, it may be advantageous to determine the homology (ie, percent identity) between the selected frame and the frame surrounding the atypical CDR to be incorporated into the selected frame. If primers with a known sequence are used to "fish" for CDRs in a library, it is of course more important to determine the homology between the primers and the frame sequence.

The degree of homology determines the number of PCR amplification steps necessary to bring the atypical CDR into the selected frame. This means that lower degrees of homology will result in more sequential PCR steps to convert the original FRs flanking the atypical CDR to the sequence of the chosen FR.

Example IA

Figure 1 shows the sequences and steps involved in the amplification of DP-29 CDRH2, and its incorporation into the nucleic acid encoding the frame of DP-47.

Part A shows nucleic acid sequences encoding parts of the framework regions flanking the DP-47 and DP-29 CDRH3 loops and the deduced amino acid sequences. Nucleotide matches are marked with the symbol I. As will be seen, there are some mismatches: respectively 8 of

36 nucleotides and 7 of 27 nucleotides, in the two flanking parts.

Part B shows amplification primers ("#primers") identical to the nucleic acid coding portions of the framework regions flanking the DP-29 CDRH2 loops, flanked by the double-stranded DP-29 coding sequence.

Part C shows the amplification product ("#lproduct") of the first PCR step (which was shown in part B). The conditions for amplification are as for CDR amplification in WO 98/32845. The #1 product is identical to the coding sequence of DP-29. Alongside this are primers ("#2primers") for a second PCR step. These are identical to the nucleic acid encoding corresponding parts of the framework regions flanking the DP-47 CDRH2 loops. As a consequence of this, the same misfits are visible as in part A.

Part D shows the product ("#2 product") of the second PCR step. This has the framework regions of DP-47 (main frame) and the CDRH2 loop of DP-29.

There is therefore sufficient sequence identity between the frame regions of DP-47 and DP-29 that flank the CDRH2 loop so that the loop can be switched from one frame to the other in a single PCR step.

The DP-29 germline gene encodes a CDHR2 of canonical class 4 (VBASE), while the CDRH2 of DP-47 is of canonical class 3 (VBASE).

The second PCR step can be performed as an overlap expansion PCR step, as the primer used is identical to the main framework sequence by which the CDR is intended to be incorporated, for example using conditions (and other primers) discussed in WO 98 /32845.

Example IB

An iterative process of sequential PCR amplifications is used to insert a CDR into a DP-47 mainframe from a germline gene (DP-73) that has significantly different sequences encoding the portions of the framework regions flanking the CDR. In this example, the homology between the DP-47 VH frame, adjacent CDRH2, and the DP-73 frame is too low to allow direct amplification (eg, in an overlap-expansion PCR step) using primers that are completely identical to DP -47. Several individual PCR steps are therefore used, each step using a unique primer pair. The primers are successively modified to become more homologous to the DP-47 primer.

In this process, it is important to choose the appropriate distribution of the base modifications carefully. Figure 2 shows this process. The underlined sequence is where the biggest differences between DP-47 and DP-73 are. Bold letters indicate residues in the primers that are identical to those in DP-47, main frame.

Parts A and B are analogous to the same parts in Figure 1. Again, there are mismatches between the sequences encoding the parts of the frame flanking the DP-47 and DP-73 CDRH2 loops, respectively 13 mismatches out of 42 nucleotides and 9 mismatches out of 27 in the two flanking sequences.

In part C, primers that are chimeric of DP-47 and DP-73 are used instead of using primers that are identical to DP-47, to introduce changes in the in-frame regions of the amplified DP-73 fragment, to bring them partially into conformity with those of DP-47. So, instead of there being no mismatches between the primers and the DP-47 sequences (as in Example IA), there are still some mismatches, though less than before, ie 2 in each flanking sequence.

Part D shows the amplification product ("#2 product") of the second PCR step, aligned with primers ("#3 primers") identical to corresponding parts of the DP-47 frame. As for example IA, such primers can be used in overlap-expansion PCR. The third amplification step (analogous to the second in Example IA) leads to a fragment incorporating CDRH2 of DP-73 into a frame of DP-47.

In particular, in the second PCR step (which uses the #2 primer, shown in part C), the following base substitutions were made in the upper PCR primer, relative to the #1 primer used in the first PCR step (shown in part B): In position 35 (calculated from the 5' end of the primer) G was changed to C; in position 37, A was changed to G and in position 38, T was changed to C. This results in a higher degree of homology than if a primer homologous to DP-47 had been used directly in the second PCR step.

The PCR intermediate from the second PCR step therefore still contains bases that are homologous to the DP73 sequence (G36 and C39). These bases will therefore not be primed in the third PCR step, as the upper primer used in this step is 100% homologous to the DP47 sequence. However, bases 34, 37 and 38 of the amplification product ("#2 product) from the second amplification step are now homologous to the DP-47 sequence and this homology will give a pairing of 6 of 8 bases at the 3' end (underlined) of the upper primer in the third PCR step (shown in part D). This should be compared to 3 of 8 bases at the 3' end between DP47 and DP73. Such an increase in homology greatly facilitates the successful production of a DNA sequence which includes the DP-73 CDRH2 in the DP-47 frame.

Using the principles of this procedure, any CDR can be transferred to any given and chosen framework resulting in composite antibody molecules that contain combinations of natural CDR loops, and thus can also have canonical structures that do not occur naturally. Combinations of atypical but natural CDR loops therefore provide a basis for generating an enormous variety in the antibody combination site and the variants formed can be captured in large libraries, e.g. using phage (Marks et al, 1991), ribosome (Hanes and Pluckthun, 1997) or covalent (WO 98/37186) display technologies.

References

Adkins JC, Spencer CM (1998) Drugs 56(4):619-26; discussion 627-8.

Al-Lazikani B, Lesk AM, Chothia C (1997) JMol Biol 273(4):927-48

Chothia C, Lesk AM (1987) JMol Biol 186 : 651-663

Chothia C et al. (1987) JMol Biol 196 : 901-917

Chothia C et al. (1989) Nature 342: 877-883

Chothia C, Lesk AM, Gherardi E, Tomlinson IM, Walter G, Marks JD, Llewelyn MB, Winter G (1992) JMol Biol 227(3):799-817

D'haens G, Van Deventer S, Van Hogezand R, et al. (1999) Gastroenterology 116(5):1029-34 Duenas M, Borrebaeck CA (1994) Biotechnology (NY) 12(10):999-1002

Griffiths AD et al. (1993) EMBO J 12 (2): 725-734

Griffiths AD et al. (1994) EMBO J 13 (14): 3245-3260

Griffiths AD and Duncan AR (1998) Curr Opin Biotechnol 9:102-108

Hanes J and Pluckthun A (1997) Proe Nati Acad Sei USA 94 (10): 4937-42

Hoogenboom HR, Winter G (1992) JMol Biol 227(2):381-8

Jirholt P, Ohlin M, Borrebaeck CAK, Soderlind E (1998) Gene 215 (2): 471-476

Kobayashi N, Soderlind E, Borrebaeck CAK (1997) Biotechniques 23 (3): 500-503

Lara-Ochoa F, Almagro JC, Vargas-Madrazo E, Conrad M (1996) Mol envol 43: 678-684

Larrick JW, Danielsson L, Brenner CA, et al. (1989) Bio/Technology 7: 934

Malmborg AC, Duenas M, Ohlin M, et al. (1996) Immunol Methods 198 (1): 51-7

Malmborg AC, Soderlind E, Frost L, Borrebaeck CA (1997) J Mol Biol 273 (3): 544-51

Marks JD, Hoogenboom HR, BonnertTP, et al. (1991) JMol Biol 222(3):581-597

McLaughlin P, White CA, Grillo-Lopez AJ, Maloney DG (1998) Oncology (Huntingt) 12(12): 1763-9; discussion 1769-70, 1775-7

Morea V, Tramontano A, Rustici M, Chothia C, Lesk AM (1998) JMol Biol 275: 269-294

Nissim A, Hoogenboom HR, Tomlinson IM, et al. (1994) EMBO J13: 692-698

Palares et al. (1999) Exp Clin. The immunogen. 16: 36-60

Pini A, Viti F, Santucci A, Carnemolla B, et al. (1998) J Biol Chem 273 (34): 21769-21776 Sheets MD, Amersdorfer P, Finnern R, et al. (1998) Proe Nati Acad Sei USA 95 (11): 6157-6162

Soderlind et al. (1999) Immunotechnology 4, 279-285

Tomlinson IM, Cox JP, Gherardi E, et al. (1995) EMBO J 14 (18): 4628-38

Vargas-Madrazo E, Lara-Ochoa F, Almagro JC (1995) JMol Biol 254:497-504

Vaughan TJ, Williams AJ, Pritchard K, et al. (1996) Nat Biotechnol 14 (3): 309-314

Williamson RA, Burioni R, Sanna PP, et al. (1993) Proe Nat Acad Sci USA 90: 4141-4145

Zebedee SL et al. (1992) Proe Nati Acad Sei USA 89 : 3175-3179

Claims (18)

1. Method for the production of a polynucleotide sequence encoding an antibody variable region, the variable region comprising complementary determining regions (CDRs) located within a selected frame ("master frame"), characterized by comprising the steps: a) obtaining of at least one nucleic acid molecule encoding one or more CDRs and associated framework regions ("original framework"); b) amplification of at least one CDR-encoding portion of the nucleic acid molecule(s) in step (a) using one or more oligonucleotide pairs as amplification primers, and c) assembly of a polynucleotide sequence encoding a antibody variable region by combining the amplified CDR-encoding nucleotide sequences produced in step b) with the nucleotide sequences encoding the main frame, where the oligonucleotide primers in step b) comprise nucleotide sequences that are different from the corresponding nucleotide sequences that code for the main frame. 2. Method for producing a library of polynucleotide sequences each encoding an antibody variable region comprising complementary determining regions (CDRs) located within a common selected frame ("master frame"), characterized by comprising the steps: a) obtaining of a population of nucleic acid molecules encoding one or more complementarity-determining regions (CDRs) and associated framework regions (the "original framework"); b) amplification of at least one CDR-encoding part of the nucleic acid molecules in step a) using one or more oligonucleotide pairs as amplification primers, and c) assembly of a polynucleotide sequence encoding an antibody variable region by combining the amplified CDR encoding nucleotide sequences produced in step b) with the nucleotide sequences encoding the main frame, where the oligonucleotide primers in step b) comprise nucleotide sequences that are different from the corresponding nucleotide sequences that code for the main frame. 3. The method in accordance with claim 1 or 2, characterized by comprising the steps i) obtaining at least one oligonucleotide pair, ii) using the oligonucleotide pair as amplification primers to amplify nucleotide sequences that code for different CDRs, and iii) assembly of polynucleotide sequences encoding antibody variable regions by incorporating nucleotide sequences derived from step ii) above, with nucleotide sequences encoding the main frame, where the oligonucleotides in step i) have sequences that are different from corresponding sequences that code for the main frame. 4. Method in accordance with claim 1 or 2, characterized in that the polynucleotide sequence(s) assembled in step c) codes for an immunoglobulin G (IgG) variable region, an IgG heavy chain or light chain, a non-naturally occurring antibody variable region, an antibody variable region comprising at least one CDR with a canonical structure unusual for CDRs in naturally occurring antibody variable regions comprising the main frame, or an antibody variable region comprising at least one CDR derived from a germline gene family other than the main frame. 5. Method in accordance with claim 1 or 2, characterized in that step a) comprises obtaining a population of nucleic acid molecules that each code for an antibody variable region. 6. Method in accordance with claim 5, characterized in that the nucleic acid molecules each code for an antibody variable region from the same germ cell line gene family. 7. Method in accordance with claim 5, characterized in that the oligonucleotide primer pairs in step b) selectively hybridize with a targeted sub-population of the nucleic acid molecules that were obtained in step a). 8. Method in accordance with claim 7, characterized in that the targeted sub-population of nucleic acid molecules each encode an antibody variable region from the same germ cell line gene family. 9. Method according to claim 6 or 8, characterized in that the nucleic acid molecules each code for an antibody variable region which is derived from the same germ cell line gene. 10. Method in accordance with claim 9, characterized in that the germ cell line gene is selected from the group comprising DP-29 and DP-73. 11. Method in accordance with claim 1 or 2, characterized in that the main frame is derived from a germ cell line gene selected from the group comprising DP-47 and DPL-3. 12. Method according to one of claims 1-11, characterized by comprising a further step, carried out after step b) but before step c), for modifying the nucleotide sequence of the amplified CDR-encoding molecules in step b), such that the parts of the amplified molecules which code for frame regions share greater sequence identity with the corresponding parts of the main frame. 13. Method in accordance with claim 12, characterized in that the nucleotide sequence of the amplified CDR-encoding molecules in step b) is modified so that the parts of the amplified molecules that code for frame regions share 100% sequence identity with the corresponding parts of the main frame. 14. Method according to claim 12 or 13, characterized in that the further step comprises a single round of PCR amplification using oligonucleotide primers comprising a nucleotide sequence which is a chimera of the nucleotide sequence which codes for the original frame and main frame. 15. Method according to claim 12 or 13, characterized in that the further step comprises carrying out one or more rounds of PCR amplification of the CDR-encoding nucleic acid molecules produced in step b), each additional round of amplification is carried out using oligonucleotide primers comprising a nucleotide sequence that has an increased number of nucleotide mismatches (eng. mismatches) in relation to the original frame sequence, the mismatches share sequence identity with the corresponding nucleotides in the main frame. 16. Method in accordance with claim 1 or 2, characterized in that step c) includes the use of overlap expansion PCR. 17. Method according to claim 1 or 2, characterized by comprising a further step for inserting the polynucleotide sequence assembled in step c) into an expression vector, preferably a phage display vector. 18. Method according to one of claims 1-17, characterized by further comprising the step of expressing the polynucleotide sequence(s) assembled in step c) and screening the resulting polypeptides, comprising an antibody variable region, for the desired properties .